Computer Science student at Ashesi (Class of 2027). I turn complex PDFs into useful AI tools, messy data into SQL dashboards, and contribute to multilingual LLM evaluation with Global PIQA.
Explore each case study, then expand Project breakdown for subproject descriptions and submitted notebooks or reports. Pfizer and Breaking Games include verified work and final deliverables.
Breaking Games · Jul 2026 · ✅ Verified by Extern
Turn six messy CSVs into an analytics database and a Q4 dashboard that drives real business decisions. SQL + Claude.
Pfizer · Oct 2026 · ✅ Verified by Extern
Prototype AI-powered document intelligence with Pfizer—using OCR, LLMs, and RAG to automate real enterprise PDF workflows and build a standout portfolio project.
Global PIQA Research Collaboration
Contributed to Global PIQA, a commonsense benchmark spanning 141 language varieties. My work focused on prompt preparation, filtering, formatting and cultural review for the Zarma dataset, in a collaboration of 350+ researchers from 65+ countries.
National Geographic Society + The Nature Conservancy
During my September-November 2025 conservation externship with National Geographic Society and The Nature Conservancy, I used GIS and spatial analysis to investigate environmental risks and biodiversity patterns. My research brief focused on restoring the Tanghin Dam ecosystem, connecting geospatial
Beats by Dre via Extern
During my June-July 2024 consumer insights externship with Beats by Dre, I explored consumer reviews using Python. The aim was to understand how customers described their experiences and turn unstructured review text into interpretable findings about consumer sentiment and behavior.
Computer Science student at Ashesi (Class of 2027). I turn complex PDFs into useful AI tools, messy data into SQL dashboards, and contribute to multilingual LLM evaluation with Global PIQA.
My work connects document intelligence, SQL analytics and multilingual NLP. Through Pfizer and Breaking Games externships, I built PDF extraction/RAG workflows and a SQLite/Chart.js dashboard. I contributed prompts, filtering and cultural review for Global PIQA's Zarma dataset. I study Computer Science at Ashesi (June 2027), am a Mastercard Foundation Scholar and Dean's List student, and completed an engineering exchange at CESI, France, with a French Government Scholarship. I also tutor Python and support Youths4Change initiatives. I work in French and English.
Externships
SQL & Database Architecture Externship with Breaking Games
Completed · Jul 2026
Experience
AI Document Intelligence Extern
Pfizer via Extern · July 2026 – October 2026
SQL & Database Architecture Externship
Breaking Games · July 2026 - September 2026
Marine Conservation Externship
National Geographic Society + TNC · September 2025 - November 2025
Consumer Insights Data Analytics Externship
Beats by Dre · June 2024 - July 2024
Founder
Association Neb lã Bũmbu · Feb 2023
Education
Ashesi University, Ghana
BSc Computer Science · Class of 2027
CESI Graduate School of Engineering, France
Digital Sciences Engineering exchange (2025-2026) · Class of 2026
Google via Coursera
Data Analytics Professional Certificate · Class of 2025
Skills
✅ Verified by Extern
Turn six messy CSVs into an analytics database and a Q4 dashboard that drives real business decisions. SQL + Claude.
Unified six e-commerce and marketing sources in SQLite (Jul-Sep 2026) to guide Q4 inventory and advertising decisions. Built dimension/fact tables, clean_checkouts and dashboard_metrics, then a Chart.js dashboard with five KPIs. Expand Project breakdown below to inspect each stage and its submitted Holiday Game Plan report.
Corrected a row-level filter (450 flagged sessions) by aggregating to cart level (269 suspect sessions). Re-estimated abandonment from 91.9% to 37.6%: a measurement correction in the provided data. Delivered five KPIs, a product leaderboard, channel charts and Q4 recommendations; documented checkout/order-ID attribution limits.
✅ Verified by Extern
Prototype AI-powered document intelligence with Pfizer—using OCR, LLMs, and RAG to automate real enterprise PDF workflows and build a standout portfolio project.
Built a Python prototype for pharmaceutical PDFs (Jul-Oct 2026), combining OCR, field extraction, bundled-document routing and a Gradio RAG assistant with supporting passages and page references. Explore Project breakdown below for the methods, notebooks, evaluation reports and UI presentation. The final integrated notebook is linked above the accomplishments.
.jpg)
Compared three OCR engines and three RAG retrieval configurations; built JSON outputs for document boundaries and types. Delivered a Gradio assistant with page-referenced answers, chat export and session state. Packaged the end-to-end prototype with a notebook, evaluation presentation and demo, documenting accuracy, latency and remaining limitations.
Global PIQA Research Collaboration
Contributed to Global PIQA, a commonsense benchmark spanning 141 language varieties. My work focused on prompt preparation, filtering, formatting and cultural review for the Zarma dataset, in a collaboration of 350+ researchers from 65+ countries.
Contributed to Global PIQA, a commonsense benchmark spanning 141 language varieties. My work focused on prompt preparation, filtering, formatting and cultural review for the Zarma dataset, in a collaboration of 350+ researchers from 65+ countries.
Prepared prompts, filtered and formatted examples, and reviewed French seed examples for cultural relevance while helping build the Zarma dataset with Mamadou K. Keita. Contributed to a collaboration of 350+ researchers from 65+ countries.
Accepted as a poster at NeurIPS 2026, Evaluations & Datasets Track. The best tested model scored 79% on Zarma, highlighting an evaluation gap in this low-resource language. My contribution focused on dataset quality and cultural relevance. Paper: https://arxiv.org/abs/2510.24081
National Geographic Society + The Nature Conservancy
During my September-November 2025 conservation externship with National Geographic Society and The Nature Conservancy, I used GIS and spatial analysis to investigate environmental risks and biodiversity patterns. My research brief focused on restoring the Tanghin Dam ecosystem, connecting geospatial
During my September-November 2025 conservation externship with National Geographic Society and The Nature Conservancy, I used GIS and spatial analysis to investigate environmental risks and biodiversity patterns. My research brief focused on restoring the Tanghin Dam ecosystem, connecting geospatial evidence to practical conservation recommendations.
Cleaned and explored geospatial data, produced maps and visualizations, and created an ArcGIS StoryMap that brings the analysis together. Prepared a concise research brief to communicate key patterns and proposed restoration actions to a wider audience.
Delivered an ArcGIS StoryMap and a research brief focused on Tanghin Dam. The project strengthened my ability to turn spatial data into a clear narrative for conservation decisions. StoryMap: https://arcg.is/08qTvP0
Beats by Dre via Extern
During my June-July 2024 consumer insights externship with Beats by Dre, I explored consumer reviews using Python. The aim was to understand how customers described their experiences and turn unstructured review text into interpretable findings about consumer sentiment and behavior.
During my June-July 2024 consumer insights externship with Beats by Dre, I explored consumer reviews using Python. The aim was to understand how customers described their experiences and turn unstructured review text into interpretable findings about consumer sentiment and behavior.
Performed sentiment analysis and exploratory data analysis, identified consumer behavior trends, and built visualizations to communicate the findings. The work combined text analysis with clear presentation for a brand and consumer-research audience.
Produced data visualizations and consumer insights to support strategic brand decision-making. The project developed my ability to analyse review text, explain patterns and communicate useful findings for brand decisions.
